Long Term Evolution (LTE) networks provide connectivity using a base station, such as an eNodeB, and a core network containing a serving gateway (SGW) and a packet gateway (PGW). The PGW is an integral part of the network that acts as a gateway to the network operator's network and also to the public Internet. However, only a small number of PGWs is typically supported by an LTE network. For example, three PGWs may support all operations of a mobile operator of a medium-size country. In certain situations, such as in a border region between PGW coverage areas, it may be desirable to have more than a small number of PGWs, and to permit handovers between these PGWs.
The PGW is the anchor node for IP services, including IP based mobility, IP address management, and other operator services such as data traffic monitoring for the purposes of QoS control, charging, etc. While an operator may maintain multiple PGWs, these gateway functions are used to provide a particular service to the whole operator network, or at least a significant portion of it. For example, an USA-based operator may maintain separate PGWs for its New England operations and its California operations, but it would not do so for each neighborhood in the city of Boston. As such, IP mobility between PGWs is not supported. In fact, the UE cannot even keep the same IP address if it changes PGWs. In the event that the UE does move from one PGW to another, it will experience an interruption of data services. Of course, given the large-scale coverage of PGWs, this is unlikely and thus lack of mobility support across PGWs is not a major concern for operators. (Zú{umlaut over (n)}iga et al., “Distributed Mobility Management: a Standards Landscape,” IEEE Communications Magazine 51(3):80-87, March 2013.)
However, the fact that a central gateway node anchors a significant amount of traffic for network's users in a very large geographical area does raise other serious issues, as introduced in Section II. Two issues are particularly acute: i) the sheer amount of traffic “hitting” the PGW; and ii) the often highly sub-optimal routes which the need to anchor traffic at the PGW forces. It would appear that one way to address this problem would simply be to allow PGWs to cover much smaller areas and/or become much more service specific. However, this would necessitate support of one key feature currently not supported by the PGW: inter-PGW mobility. (Id.)